Skip to main content
. 2020 Nov 12;10:19622. doi: 10.1038/s41598-020-76164-5

Table 3.

Estimated range reduction and expansion based on differences between presD-L and pastD-L evaluated based on six thresholds. presD-L is estimated present distribution using average climate data and present land cover data, and pastD-L is estimated past distribution using average climate data and past land cover data. Mean of the area (M) and the percentage of difference (Δ) were calculated using 10 replicates. The area is the number of presence cells (1 km2) in pastD-L.

B. diversus B. ardens B. hypocrita B. ignitus B. honshuensis B. beaticola
Minimum training presence (MTP)
M 360,567.8 350,768.2 341,197.8 273,290.1 117,315 54,568.2
Δ − 0.365% 1.14% 0.924% 0.551% 6.03% 1.47%
10 percentile training presence (10PTP)
M 192,053.6 170,564.9 125,860 87,775.6 22,206.5 13,082.4
Δ − 3.38% 4.73% 0.971% 4.35% 6.68% 5.13%
Equal training sensitivity and specificity (ETSS)
M 85,870.7 91,713.2 58,770.3 53,167.7 20,659.8 16,688.3
Δ − 0.754% 11.3% 4.09% 7.07% 6.44% 4.54%
Maximum training sensitivity plus specificity (MTSS)
M 94,149.7 106,783.9 64,916.6 50,873.3 25,319.7 21,899.9
Δ − 1.19% 9.22% 3.61% 7.38% 7.16% 4.08%
Balance training omission, predicted area and threshold value (BTPT)
M 340,052.7 263,560.4 233,491.3 157,102.4 117,315 62,099.6
Δ − 0.899% 2.59% 2.05% 1.69% 6.03% 0.978%
Equate entropy of thresholded and original distribution (EETO)
M 229,879.7 173,350.1 141,685.7 90,857.7 38,965.9 19,011
Δ − 2.86% 4.68% 1.37% 4.17% 7.19% 4.27%